While virtually all human tumors are derived from a single cell-of-origin, neoplastic cells within a tumor evolve over time due to genetic and epigenetic alterations, lineage diversification, and influences from stromal cells. These processes generate considerable heterogeneity within populations of neoplastic cells from individual tumors. Despite our knowledge of the existence of tumor cell heterogeneity, it is not understood whether the heterogeneous subpopulations of tumor cells merely co-exist, or alternatively whether they communicate with each other, complementing one another's phenotypes and generating biological outcomes that individual populations are incapable of producing on their own. The lack of understanding of such functional interactions between tum9r populations has been due in large part to the inability to maintain the heterogeneity of human tumors in culture and to propagate distinct clonal subpopulations from individual tumors. We and our collaborators have developed approaches that have overcome these barriers and made it feasible to culture bar-coded, fluorescently-tagged, clonal populations of tumor cells and then track individual clonal populations within tumor xenografts generated from mixtures of transplanted clones. Using these approaches, we have obtained evidence the supports the hypothesis that clonal subpopulations within a tumor cooperate with one another to promote tumor expansion and metastasis. In this proposal, we describe plans to test this hypothesis in human breast tumors by (1) characterizing the extent of genetic and phenotypic variation among clonal populations derived from an individual breast tumor and analyzing the tumor-initiating, invasive, and metastatic activity of each clonal subpopulation, (2) generating a map that plots the localization of clonal populations within tumors and their dynamic evolution over time, (3) investigating the functional consequences of heterogeneity within human breast tumor cell populations, and (4) elucidating the mechanisms responsible for phenotypes generated by intratumoral crosstalk. These studies will provide important insights into the nature of cooperative interactions between tumor cell populations and how these affect tumor expansion, invasion, or metastasis.